Re: How do scienctists know which traits are dominant or recessive in humans?

People once believed that this is a very straightforward issue; brown eye 
color is dominant, and blue eye color is recessive. The question has grown 
more complicated (it now looks like more than one gene controls eye 
color), but the original conclusion is mostly right, it seems.

How do we know? Well, a traditional dominant allele has the following 
behavior, if we assume perfect penetrance. (Perfect penetrance means that 
anybody who carries the allele shows its associated phenotype.)

1) If a person has the trait, then at least one of their parents must 
always have the trait. This is because only one copy of a dominant allele 
is needed for the phenotype to show, and the person had to get it somehow.

2) If a parent has the dominant trait, then there is at least a fifty-
percent chance that each of his or her children will have the trait. Any 
of the children that have the trait have a fifty-percent chance of passing 
it on to one of their children, unless the mother also had the trait, in 
which case the odds could be higher.

When we see a trait exhibiting the above behavior, as brown eyes do, we 
have good reason to suspect that it is dominant.

Recessive alleles, in contrast, have different behavior:

1) Somebody can only have the trait if both parents carried at least one 
copy each of the recessive allele. They did not need to show the trait, as 
one copy isn't enough to reveal it, but they must both carry it. As a 
result, recessive traits skip generations in many cases. Grandparents and 
their grandchildren may well show it even if the intervening parents did 
not.

2) If a parent has the trait and mates with somebody who does not carry 
the gene, ALL of their children will lack the trait. 

The blue eyes trait has these characteristics, so it is recessive.

Recessive traits are a bit harder to deal with, in many ways, because 
carriers of the genes often don't show the trait. Here, statistics help 
tell apart recessives from dominants and semidominants with low penetrance.

Although observing eye color over multiple generations and looking at the 
pedigrees leads to the conclusion that brown eyes are dominant to blue 
eyes, one can verify this by applying statistical methods, such as a 
likelihood ratio test, to a large number of pedigrees.

In this test, one assumes the gene is recessive, then calculates a 
likelihood score representing the chance of seeing the pattern of traits 
in the pedigrees based on that assumption. One then changes the assumption 
to one of dominance and calculates a likelihood score of the results shown 
if dominance is assumed. Finally, one compares the two scores to see if 
one is significantly larger; this will suggest dominance or recessivity. 
This is not a perfect test, in that it doesn't directly take into account 
the preponderance of dominant or recessive alleles (though it can be 
modified to do that), but it works.

I hope that helped.

Paul Nagami

ADDITIONAL SOURCE:
 http://www3.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=227240

Briefly discusses a linkage test.